IPB

Welcome Guest ( Log In | Register )

2 Pages V   1 2 >  
Reply to this topicStart new topic
Upsample 44.1 24-bit to 48 24-bit for DVD
LTD
post Sep 17 2013, 23:01
Post #1





Group: Members
Posts: 7
Joined: 16-September 13
Member No.: 110117



I have some 44.1 24-bit flac files that I want to put on dvd so I need to upscale to 48. Currently I could use Adobe Audition and Iíve also used eac3to in the past. Is there a difference in quality depending on the program used? If so what are the best ways to do it? And what are the best free programs for this sort of thing? Thanks.
Go to the top of the page
+Quote Post
DVDdoug
post Sep 17 2013, 23:35
Post #2





Group: Members
Posts: 2441
Joined: 24-August 07
From: Silicon Valley
Member No.: 46454



I've upsampled & downsampled going back-and-forth between CD and DVD, and I've NEVER heard any difference no matter what application I was using. (I mostly use GoldWave, which used to have a "poor" resampler, but it's since been improved.)

SOX (command line) is known to have a good resampler. I'm sure Audition is audibly as good as anything else.
Go to the top of the page
+Quote Post
Arnold B. Kruege...
post Sep 18 2013, 03:46
Post #3





Group: Members
Posts: 3535
Joined: 29-October 08
From: USA, 48236
Member No.: 61311



QUOTE (LTD @ Sep 17 2013, 18:01) *
I have some 44.1 24-bit flac files that I want to put on dvd so I need to upscale to 48. Currently I could use Adobe Audition and Iíve also used eac3to in the past. Is there a difference in quality depending on the program used?


There are good upsamplers and there are bad ones. The good ones in my book behave in accordance with an theoretical ideal upsampler.

A theoretical ideal upsampler has no effect on sound quality.

But since you seem to think that upsampling is going to *improve* sound sound quality, you want what I would call a bad one.

I think that Iztope Ozone includes some bad upsamplers. But, it costs money.
Go to the top of the page
+Quote Post
jensend
post Sep 18 2013, 06:41
Post #4





Group: Members
Posts: 140
Joined: 21-May 05
Member No.: 22191



Arnold, I don't see any signs that LTD is asking for magic here. On the contrary, it appears he's resampling to fit the DVD format, not because he thinks upsampling will improve things.

LTD, recent versions of Audition (CS 5.5+), eac3to, SoX, and many other applications should all give perceptually indistinguishable results.

If you want to get into details: a resampler faces tradeoffs between filter complexity, aliasing, transition band width, ripple, pre-echo, etc. Once you've learned what those terms mean (try the Hydrogenaudio wiki or Wikipedia) you can check out the Sample Rate Converter Comparisons page to get an idea how some programs perform. Keep in mind that since their scales go down to -180dB some differences which are very visible on these graphs are actually inaudible.

A dozen years ago, a lot of programs used very simple filters and/or naive algorithms, with adequate speed on slow computers but with negative impacts on audio quality. Linear or ZOH resampling would be extreme cases; you can find those as an option under Secret Rabbit Code at the comparisons page. These days processor power is in ready supply and many programs use filters that are kind of overkill; SoX VHQ linear phase is a definite example of that.
Go to the top of the page
+Quote Post
bandpass
post Sep 18 2013, 08:00
Post #5





Group: Members
Posts: 321
Joined: 3-August 08
From: UK
Member No.: 56644



CODE
sox in.flac out.flac rate 48k
is a good baseline. As jensend says, there are tradeoffs, and you can make these by adding command-ine options if you want, in particular if you believe your files, playback system, or ears are in some way unusual.

E.g. you say that the files are 24-bit: if (possible but unlikely) there is useful information in the bottom 8 bits, then this can be preserved using
CODE
sox in.flac out.flac rate -v 48k
(VHQ mode as mentioned above); the downside of this is slightly longer processing time and a longer filter. Is a longer filter bad? Often not, but one situation in which it may be worse is segues: if you have segues then you should concatenate, resample, then resplit the segued files to avoid potential problems completely.

You can use a shell for-loop for batch processing, or foobar2000.

Go to the top of the page
+Quote Post
LTD
post Sep 19 2013, 03:59
Post #6





Group: Members
Posts: 7
Joined: 16-September 13
Member No.: 110117



Thanks for the replies. I donít have a home internet connection so I canít always respond quickly. I am definitely not trying to increase the sound quality, but I donít want to degrade the quality with a resample. Since they are 24-bit I want to burn them to DVD. I just want to play these with my bluray connected to my stereo, instead of playing them through my sound card. I have Audition 1.5 and 3.0. I could download a trial of Audition CS. Iím willing to try something else too.
QUOTE
E.g. you say that the files are 24-bit: if (possible but unlikely) there is useful information in the bottom 8 bits, then this can be preserved using

These files are the Beatles 24-bit Flac thumbdrive. When you say the bottom 8 bits, is this the extra information compared to the 16-bit releases (CDs)? I sure hope thereís a difference between the two! After the responses I got from my other thread, who knows if Iíll actually be able to tell the difference.

QUOTE
(VHQ mode as mentioned above); the downside of this is slightly longer processing time and a longer filter. Is a longer filter bad? Often not, but one situation in which it may be worse is segues: if you have segues then you should concatenate, resample, then resplit the segued files to avoid potential problems completely.

Are segues when one track leads seamlessly to the next with no silence in between? If so then that is a very important consideration to me.
Go to the top of the page
+Quote Post
TomasPin
post Sep 19 2013, 04:47
Post #7





Group: Members
Posts: 168
Joined: 5-June 13
From: Argentina
Member No.: 108508



QUOTE (LTD @ Sep 18 2013, 23:59) *
Since they are 24-bit I want to burn them to DVD.

Why? You can downsample to 16-bits and burn them into a CD just as well. It will sound the same.

QUOTE
I sure hope there’s a difference between the two!

If there is, it will be due to different mastering jobs - the 24-bit version done with less dynamic range compression, limiting, and so forth. They want you to think it sounds better because the files are "hi-rez". Not true.

Sorry if I sound harsh, but that's how it happens.

This post has been edited by TomasPin: Sep 19 2013, 04:49


--------------------
A man and his music: http://tubular.net/
Go to the top of the page
+Quote Post
bandpass
post Sep 19 2013, 07:40
Post #8





Group: Members
Posts: 321
Joined: 3-August 08
From: UK
Member No.: 56644



QUOTE (LTD @ Sep 19 2013, 03:59) *
Are segues when one track leads seamlessly to the next with no silence in between?

Yes. You might get away without concatenating but personally, I wouldn't take the risk (of hearing a small click at the join). Considerations include preserving meta data and precise track start times (if you want to be able play segued tracks individually). I don't know if there's an easy, 'off the shelf' way of doing this -- I've used ad hoc shell-scripts in the past.

Without resplitting it's relatively straight-forward, e.g.:
CODE
sox 0[1-8]*.flac side1.wav rate -v 48k
Go to the top of the page
+Quote Post
Wombat
post Sep 19 2013, 14:54
Post #9





Group: Members
Posts: 950
Joined: 7-October 01
Member No.: 235



QUOTE (bandpass @ Sep 18 2013, 09:00) *
VHQ mode as mentioned above); the downside of this is slightly longer processing time and a longer filter. Is a longer filter bad? Often not, but one situation in which it may be worse is segues: if you have segues then you should concatenate, resample, then resplit the segued files to avoid potential problems completely.

Interesting information, thanks. I stumbled across that problem but never thought it only happens because of that. I saw Korg for example does a very short fade-in and out on the dsd to pcm conversion. This avoids clicks but can by itself make some short stuttering sound. What about an option that calculates the very first samples only at lower precision?
Go to the top of the page
+Quote Post
pdq
post Sep 19 2013, 15:59
Post #10





Group: Members
Posts: 3305
Joined: 1-September 05
From: SE Pennsylvania
Member No.: 24233



QUOTE (Wombat @ Sep 19 2013, 09:54) *
What about an option that calculates the very first samples only at lower precision?

It is also possible to come up with a set of coefficients that results in a filtered value, not of the central point, but of a point at one end or the other of a group of data points.
Go to the top of the page
+Quote Post
LTD
post Sep 19 2013, 23:47
Post #11





Group: Members
Posts: 7
Joined: 16-September 13
Member No.: 110117



QUOTE
Why? You can downsample to 16-bits and burn them into a CD just as well. It will sound the same.


Is this what everyone here thinks? Why would recording studios use 24-bit mastering for CDs? How do 96KHz or 192KHz sound files compare to CD?
Go to the top of the page
+Quote Post
greynol
post Sep 19 2013, 23:54
Post #12





Group: Super Moderator
Posts: 10000
Joined: 1-April 04
From: San Francisco
Member No.: 13167



QUOTE (LTD @ Sep 19 2013, 15:47) *
How do 96KHz or 192KHz sound files compare to CD?

The same unless you can prove otherwise in a properly-controlled double-blind test using a sample rate converter of our choosing.

QUOTE (LTD @ Sep 19 2013, 15:47) *
Why would recording studios use 24-bit mastering for CDs?

Do you realize that if you're trying to argue then this can be dismissed outright as a logical fallacy?

This post has been edited by greynol: Sep 19 2013, 23:55


--------------------
Your eyes cannot hear.
Go to the top of the page
+Quote Post
LTD
post Sep 19 2013, 23:57
Post #13





Group: Members
Posts: 7
Joined: 16-September 13
Member No.: 110117



I'm not trying to argue. It's a question. I'm trying to learn.

This post has been edited by LTD: Sep 19 2013, 23:57
Go to the top of the page
+Quote Post
saratoga
post Sep 19 2013, 23:57
Post #14





Group: Members
Posts: 4715
Joined: 2-September 02
Member No.: 3264



QUOTE (LTD @ Sep 19 2013, 18:47) *
Why would recording studios use 24-bit mastering for CDs?


Mastering is not the same as playback. Higher bit depths for playback allow for things like digital gain adjustment. That is quite useful for mastering, but irrelevant for playback.
Go to the top of the page
+Quote Post
greynol
post Sep 20 2013, 00:05
Post #15





Group: Super Moderator
Posts: 10000
Joined: 1-April 04
From: San Francisco
Member No.: 13167



QUOTE (LTD @ Sep 19 2013, 15:57) *
I'm not trying to argue. It's a question.

I'm used to being defensive about this stuff.

QUOTE (LTD @ Sep 19 2013, 15:57) *
I'm trying to learn.

http://xiph.org/video/


--------------------
Your eyes cannot hear.
Go to the top of the page
+Quote Post
LTD
post Sep 20 2013, 00:42
Post #16





Group: Members
Posts: 7
Joined: 16-September 13
Member No.: 110117



Sorry if my language is vague, I'm obviously new to HydrogenAudio. I'm not trying to sway anybody's beliefs or make any claims. I'm curious now about the audible difference between a CD and music with a sample rate of 96 or 192.
Go to the top of the page
+Quote Post
testyou
post Sep 20 2013, 00:51
Post #17





Group: Members
Posts: 90
Joined: 24-September 10
Member No.: 84113



QUOTE (LTD @ Sep 19 2013, 16:42) *
I'm curious now about the audible difference between a CD and music with a sample rate of 96 or 192.

You can perform a listening test of your own to determine if you can differentiate between test samples with different rates.

What is a blind ABX test ?
Listening Tests Forum

This post has been edited by testyou: Sep 20 2013, 00:55
Go to the top of the page
+Quote Post
saratoga
post Sep 20 2013, 00:59
Post #18





Group: Members
Posts: 4715
Joined: 2-September 02
Member No.: 3264



http://en.wikipedia.org/wiki/Audio_sampling_rate
Go to the top of the page
+Quote Post
greynol
post Sep 20 2013, 01:39
Post #19





Group: Super Moderator
Posts: 10000
Joined: 1-April 04
From: San Francisco
Member No.: 13167



QUOTE (LTD @ Sep 19 2013, 16:42) *
I'm curious now about the audible difference between a CD and music with a sample rate of 96 or 192.

There is no audible difference...
QUOTE (greynol @ Sep 19 2013, 15:54) *
unless you can prove otherwise in a properly-controlled double-blind test using a sample rate converter of our choosing.


This post has been edited by greynol: Sep 20 2013, 01:39


--------------------
Your eyes cannot hear.
Go to the top of the page
+Quote Post
TomasPin
post Sep 20 2013, 02:22
Post #20





Group: Members
Posts: 168
Joined: 5-June 13
From: Argentina
Member No.: 108508



QUOTE (greynol @ Sep 19 2013, 20:05) *

You know, that link should be frontpaged in big bold colored letters. I know it's under news, but I mean making it more notorious. Just an off-topic idea...

QUOTE (greynol @ Sep 19 2013, 21:39) *
QUOTE (LTD @ Sep 19 2013, 16:42) *
I'm curious now about the audible difference between a CD and music with a sample rate of 96 or 192.

There is no audible difference...
QUOTE (greynol @ Sep 19 2013, 15:54) *
unless you can prove otherwise in a properly-controlled double-blind test using a sample rate converter of our choosing.


I reckon reading frequencies that high can have a negative impact on the lower, audible ones when played back on certain hardware. 24-bit is better in that regard, pointless but harmless.

EDIT: But you would still need an ABX log to prove it.

This post has been edited by TomasPin: Sep 20 2013, 02:25


--------------------
A man and his music: http://tubular.net/
Go to the top of the page
+Quote Post
jensend
post Sep 20 2013, 03:31
Post #21





Group: Members
Posts: 140
Joined: 21-May 05
Member No.: 22191



QUOTE (greynol @ Sep 19 2013, 16:54) *
Do you realize that if you're trying to argue then this can be dismissed outright as a logical fallacy?
Good grief, Greynol. That's an uncivil response to a simple question. (It's also false.*)
QUOTE (greynol @ Sep 19 2013, 17:05) *
QUOTE (LTD @ Sep 19 2013, 15:57) *
I'm not trying to argue. It's a question.
I'm used to being defensive about this stuff.
That sounds like a personal problem.

Seriously.

If what you're "used to" and the built-up scars and resentments from battling audiophools in the past have taken so much of a toll that you can't respond to a simple question with respect, maybe you need to take a break. Take a chill pill. Get some perspective.

Already twice in this thread LTD has met with belittling and confrontative remarks. This seems to have become the rule rather than the exception at HA in recent months. This needs to change, and the moderators should be setting an example of civility.

Also, telling every newbie who comes along with a question to go spend a bunch of time trying to see whether they can ABX high sample rates or high bitdepth or whatnot is unproductive. We already have the answers to those questions, backed up by a lot more than just a single listener's experience. One time in 20, a single listener trying to ABX something will get a "statistically significant" result even when they're just guessing. Linking to the xiph video was much more productive; tell the uninformed what the accepted wisdom is, and then let them know they could verify it for themselves if they're so inclined or if they need further convincing.

*The only reason I see for your statement is that you're trying to claim it was an appeal to authority, but an appeal to an apparently legitimate authority is reasonable supporting evidence in an inductive argument, and cannot simply be "dismissed outright." If you're debating against someone who appeals to authority, legitimate strategies include telling them why the claimed authority is not actually an authority on the relevant subject, why the authorities' opinion doesn't apply to the present case, or how the authority came to be mistaken. Or you can simply outweigh the authority with better evidence. Trying to browbeat them by yelling "THAT'S A FALLACY!" is illegitimate rhetoric.
Go to the top of the page
+Quote Post
greynol
post Sep 20 2013, 05:02
Post #22





Group: Super Moderator
Posts: 10000
Joined: 1-April 04
From: San Francisco
Member No.: 13167



If you want know why I wasn't right, read the post just after mine.

I feel the relevant term now is non-sequitur.

If you wish to comment about the forum start a new discussion or send a PM. This discussion is not your soap box no more than it is mine or anyone else's. Further posts on the matter will be binned.


--------------------
Your eyes cannot hear.
Go to the top of the page
+Quote Post
Arnold B. Kruege...
post Sep 20 2013, 10:46
Post #23





Group: Members
Posts: 3535
Joined: 29-October 08
From: USA, 48236
Member No.: 61311



QUOTE (LTD @ Sep 19 2013, 19:42) *
I'm curious now about the audible difference between a CD and music with a sample rate of 96 or 192.


To this day, all reliable evidence points to the idea that if your audio system is clean, then there is no difference. Obtaining adequate proof of there being a difference is one of the Golden Rings, Lost Chalices, or Lost Arks of Audio and this has been since no later than the early 1990s. Many have tried at home, in universities, and in industrial labs. We have all failed.

It is true that if your audio system has not uncommon problems, such as audible high frequency nonlinear distortion, there will be a difference in some cases, but it has also been shown that if you make theaudible nonlinear distortion go away, so does audibility.

It actually turns out that the vast majority of modern clean wide range program material can be sampled as low as 32 KHz (16 KHz bandpass) without any ill effects. So not only is 44 KHz adequate, it is overkill. You might say "I can hear pure tones at 18 KHz" and last time I tried so can I and I'm 67. But do not underestimate the power of spectral masking.
Go to the top of the page
+Quote Post
LTD
post Sep 21 2013, 20:10
Post #24





Group: Members
Posts: 7
Joined: 16-September 13
Member No.: 110117



Thanks for the replies. I have a couple questions at this point.

First, what are some preferred ways to downsample 44.1 24-bit and 96 24-bit files to CD quality 44.1 16-bit?

Second, if there is no audible difference between CD and 96KHz or 192Khz, what does the HydrogenAudio community think of SACD, DVD-A, and HD audio tracks (96 or 192)?
Go to the top of the page
+Quote Post
saratoga
post Sep 21 2013, 20:17
Post #25





Group: Members
Posts: 4715
Joined: 2-September 02
Member No.: 3264



QUOTE (LTD @ Sep 21 2013, 15:10) *
First, what are some preferred ways to downsample 44.1 24-bit and 96 24-bit files to CD quality 44.1 16-bit?


For changing sampling rate, use a reasonably good resampler, of which there are a very large number to choose from:

http://src.infinitewave.ca/

For converting to 16 bit, in theory you should use dither, in practice it usually doesn't matter very much because the 16 bit is already very high dynamic range.

QUOTE (LTD @ Sep 21 2013, 15:10) *
Second, if there is no audible difference between CD and 96KHz or 192Khz, what does the HydrogenAudio community think of SACD, DVD-A, and HD audio tracks (96 or 192)?


DVD-A and SACD are nice because they give you 5.1 audio. The higher sample rate thing though is just pure marketing.
Go to the top of the page
+Quote Post

2 Pages V   1 2 >
Reply to this topicStart new topic
1 User(s) are reading this topic (1 Guests and 0 Anonymous Users)
0 Members:

 



RSS Lo-Fi Version Time is now: 18th April 2014 - 23:07